Enhancing feedback for clinical use: Creating and evaluating profiles of clients seeking counseling.

Abstract

The current study explored the reliability and clinical utility of a method designed to identify latent classes of students seeking counseling, based on 8 symptom domains and their interactions. Participants were over 50,000 college students in counseling, assessed with the CCAPS-62 and -34 as part of routine clinical care. Latent profile analysis was used to group an exploratory and confirmatory sample of students by reported symptoms across the 8 CCAPS subscales. Profiles were evaluated for reliability and clinical utility, in particular for risk assessment and the prediction of treatment duration and success. Nine reliably stable latent profiles, or groups of profiles, emerged from analysis. Profiles differed significantly in reported symptoms, demographic makeup, psychosocial history, and diagnoses. Additionally, profiles appeared to capture meaningful differences between clients that had implications for relative risk of suicide, self-harm, and violence toward others as well as significant differences in the number of sessions in treatment and the effect size of treatment. Latent profiles of patients appear to capture meaningful, stable differences that could be implemented in an automated system of evaluation and feedback, and that might be useful to clinicians, administrators, and researchers.